- Thread Starter
- #921
OK, fixed that and another one that was wrong.
Amazing. When you think no one is paying attention vs when slothful decisions are exposed and a company actually has to put nose to grindstone and earn their rep. Now if only Amir's review had enough influence to get Marantz to actually try instead of knocking out good enough feature laden models.
They do. But as I noted earlier, their test suite is only available to licencees, and we have no idea what standards they require. Given the generally poor state of HT offerings one can only assume that the standards are depressingly low, and geared towards selling lots of licenses. If their test suite was available, things might change. But Dolby (and DTS) have a vested interest in making the standards only just “good enough”. Despite their marketing.How is that bug even possible?
Doesn't Dolby require certification that their codecs function correctly before letting a company put double-Ds on a product?
No Dolby code is needed for my tests as I am feeding the unit uncompressed PCM. The Dolby Upmixer (for 2 channel to multi) was in the pipeline but told to not do anything since I was using Reference Stereo. Turns out the code was screwing with the bits even though it was configured to do otherwise. The fix was to bypass the module completely.How is that bug even possible?
Doesn't Dolby require certification that their codecs function correctly before letting a company put double-Ds on a product?
Dolby does not provide a sample rate converter for its upmixer, or for anything.My guess is that the upmixer has a crappy sample rate converter.
You should somehow also test AV-gear with surround processing/room correction enabled. I mean, most of the use is with those enabled and a perfect score in direct/pure mode doesn't tell what the actual performance is in regular everyday usage.My guess is that the upmixer has a crappy sample rate converter. When you tell it to pass through, it still goes through the sample rate converters but doesn't produce more channels. And hence the artifacts that we saw.
The internal algorithm could choose to do whatever it wants. It can separate the incoming stream into separate bands (at different sample rates), process and output.Dolby does not provide a sample rate converter for its upmixer, or for anything.
I hear you but there is no easy pass through mode here to test. How would you get every EQ system to be active but produce a flat response so different equipment can be tested using the same correction?You should somehow also test AV-gear with surround processing/room correction enabled. I mean, most of the use is with those enabled and a perfect score in direct/pure mode doesn't tell what the actual performance is in regular everyday usage.
I don't know... Maybe just feed Dolby or DTS encoded multichannel test signal and measure how it looks, at least that would tell if it's OK. But that would still leave EQ/Room Correction out.I hear you but there is no easy pass through mode here to test. How would you get every EQ system to be active but produce a flat response so different equipment can be tested using the same correction?
Flatness of the response is probably the least of the ills. Doing such a test would very likely uncover more nasties related to resampling in all its guises. It didn't exactly take long to turn up one such nasty. I will bet there are many more buried bodies to be found.I hear you but there is no easy pass through mode here to test. How would you get every EQ system to be active but produce a flat response so different equipment can be tested using the same correction?
Note that I cleaned up the above graph, taking out the 2.7 volt output of the HTP-1 as that was not a fair comparison against likes of RMC-1 which produce 4 volt output.
Did you ever produce a SINAD vs Measured Level chart for the HTP-1? Apologies if I missed it. I'm just curious as to where its "peak performance" area is (since it seems to have tested better at 2.7Vrms than at 4Vrms).
I also wonder if it would look any different now that they've released a firmware update for the issue, although they still have a jitter issue to figure out.
As far as I was able to determine from the mass of various posts on their forums it was a problem of the Dolby processing occurring even when it wasn't supposed to be present (like in the reference or direct mode) which was increasing noise. It wasn't all that clearly stated or maybe my brain had already gone soft from the previous 20 or so pages of posts I read, but that's what it seemed Lonnie (I think) was saying they'd found.How is that bug even possible?
Doesn't Dolby require certification that their codecs function correctly before letting a company put double-Ds on a product?